Characterization of Four Members of the Alpha-Tubulin Gene

Characterization of Four Members of theAlpha-Tubulin Gene Family in Ceratopteris richardii

RODNEY J. SCOTT

Department of Biology, Wheaton College, Wheaton, IL 60187

GERALD J. GASTONY

Department of Biology, Indiana University, Bloomington, IN 47405-7005

JEREMY W. WEATHERFORD

Department of Computer Science and Engineering, Washington University, St. Louis, MO 63108

TAKUYA NAKAZATO

Department of Biology, Indiana University, Bloomington, IN 47405-7005

ABSTRACT.—Four members of the alpha-tubulin gene family were examined in Ceratopteris

richardii. Genetic linkage mapping based on a population of nearly 500 Doubled Haploid Lines was

able to position three or four members of this gene family on linkage groups 17, 24, and 28,

respectively (two of the four observed polymorphic restriction fragments containing alpha-tubulin

genes are either identical or map too close to each other on linkage group 17 to be distinguishable

in map distance). Non-mapable monomorphic bands observed on probed Southern blots suggest

that the alpha-tubulin gene family in this species is large. Four alpha-tubulin genes from C.richardii were sequenced and found to be fairly similar to each other in terms of their amino acid

sequences, with their greatest diversity at the carboxy-terminal ends. BLAST comparisons found

each of these four amino acid sequences more similar to an alpha-tubulin from a dicot,

gymnosperm, or alga species than it was to any other alpha-tubulin sequence presently known

from Ceratopteris or from the fern Anemia phyllitidis or the moss Physcomitrella patens. Bayesian

phylogenetic analysis of nucleotide sequences placed three of the four Ceratopteris alpha-tubulin

gene copies in a clade with copies from Pseudotsuga and Anemia, consistent with a history of two

gene duplication events, one following and one preceding the divergence of ferns and seed plants.

The fourth copy is robustly separated from the preceding three and placed in a clade of algal alpha-

tubulin genes, suggesting its divergence from the ancestor of the other three before the divergence

of algae and land plants. As characterized thus far, the alpha-tubulin gene family of C. richardii is

relatively large as compared to the six copies known from fully sequenced Arabidopsis thaliana,

a condition that may be correlated with the large genome size and diverse life history constraints of

this homosporous fern species. These findings suggest several new opportunities for research into

the evolution, function, and regulation of the alpha-tubulin gene family in Ceratopteris.

This report describes the application of DNA sequencing and genetic linkagemapping to the alpha-tubulin genes of Ceratopteris richardii and shows howsuch studies can further enhance the utility of this model system. Initialsuccesses with Ceratopteris were in the areas of cytogenetics (e.g., Hickok,1976, 1977a, 1977b, 1978, 1979a, 1979b; Hickok and Klekowski, 1973, 1974),physiology and development, and Mendelian genetics (reviewed in Hickok,1987; Hickok et al., 1987, 1995). More recent studies have focused on themolecular genetics of Ceratopteris (Munster et al., 1997; Hasebe et al., 1998;Aso et al., 1999; Stout et al., 2003; Rutherford et al., 2004; Salmi et al., 2005).These recent studies that employ robust molecular methods are especiallyencouraging, since the lack of a technique to induce stable genetic trans-

American Fern Journal 97(2):47–65 (2007)

formation in Ceratopteris has undoubtedly impeded its function as a unifyingmodel system. In this report, DNA sequences from four unique alpha-tubulingenes of C. richardii are described, and the phylogenetic relationships of thesegenes to other plant alpha-tubulin genes are inferred. Molecular linkage datafor three polymorphic alpha-tubulin loci of C. richardii are used to positionthese loci on the new genetic linkage map of this species, and several potentialstrategies for using this new information in molecular studies are outlined.These perspectives should provide further encouragement for those seeking toutilize Ceratopteris as a model system.

The alpha-tubulin gene family was selected for these studies because of thebiological significance and tractability of tubulin in the gametophytegeneration of ferns and because of the new insights that may be gained forgenomics by studying them. Inferring evolutionary relationships amongmembers of gene families is generally problematic because distinguishingparalogues and orthologues is difficult without appropriate known outgroupsto polarize the gene phylogeny. Tubulin genes are ideal for studying genefamily evolution because 1) alpha-, beta-, and gamma-tubulin genes are knownto have diverged well before the divergence of plant lineages, and 2) the highlyconserved nature of the genes allows their sequences to be aligned. Thesefeatures allow inference of gene phylogenies within a given tubulin genegroup, using sequences from other tubulin groups as outgroups.

Tubulins and the microtubules they form are obviously essential compo-nents of all eukaryotic cells. In fern gametophytes, however, their roles aredirectly observable in various stages of development that can be easily studied.For example, one of the first events that signals preparation for sporegermination is the migration of the nucleus within the cytoplasm (Banks,1999). This event, which is critical to the continued development of thegametophyte, is inhibited in Onoclea sensibilis by several microtubuleinhibitors, including colchicine (Vogelmann et al., 1981). An important rolefor microtubules at a later stage of development has been suggested by thestudies of Murata et al. (1997) on blue light-induced inhibition of cell growthin dark-grown Ceratopteris gametophytes. These investigators found thatcortical microtubules reorient in response to blue light at the same time thatinhibition of cell elongation occurs. Microtubules also serve a key role in theorganization and function of fern sperm (Raghavan, 1989), and Ceratopterissperm has been used extensively in studies characterizing these roles(Hoffman and Vaughn, 1995a, 1995b, 1996; Hoffman et al., 1994; Renzagliaet al., 2004).

Because homosporous ferns compose most of the sister group to seed plants(including flowering plants; Pryer et al., 2001), increased knowledge ofgenome structure and organization in the genomes of homosporous ferns willsignificantly broaden our knowledge of genome structure and evolution invascular plants. Furthermore, characterization of the alpha-tubulin genefamily in Ceratopteris will ultimately help to answer questions specificallyrelated to the origin and function of gene families in organisms with largegenomes. At 11,294 Mb (Jo Ann Banks, personal communication), the genome

48 AMERICAN FERN JOURNAL: VOLUME 97 NUMBER 2 (2007)

of C. richardii is ca. 110 times the genome size of Arabidopsis thaliana, 24times that of rice, 12 times that of tomato, 4 times that of maize, and 2 timesthat of barley. Some relevant questions include the following. Do organismswith larger genomes have larger gene families? Do such organisms (with theirapparent excess amounts of DNA) have more pseudogenes among the membersof their gene families? How are the various members of gene families regulatedduring development in organisms with large genomes? The latter question isan especially interesting one as it relates to the ferns, whose alternation ofgenerations features fully independent sporophyte and gametophyte genera-tions.

This study employed two distinct strategies for characterizing alpha-tubulingenes from C. richardii. The first approach involved identifying, isolating, andsequencing alpha-tubulin genes that were expressed in the gametophytegeneration. This was accomplished by using an antibody-based screeningmethod with a cDNA library derived from the gametophyte generation.Bayesian Inference analyses determined the phylogenetic relationships ofthese newly obtained sequences to other plant alpha-tubulin genes. Thesecond approach utilized a mapping strategy developed for the recentlycompleted project of Nakazato et al. (2006) to generate a high-resolutionlinkage map for C. richardii. This mapping project used genetic polymorph-isms present in a population of nearly 500 Doubled Haploid Lines (DHLs)generated from an initial cross between diploid inbred lines of highly divergedgeographic races of C. richardii (Hickok et al., 1995): WN8 (derived froma Nicaraguan collection) and Ha-PQ45, a mutant of Hn-n (derived from a Cubancollection). Together these two approaches provide detailed insights regardinga previously uncharacterized gene family in C. richardii. These new insightssuggest novel strategies for studying the alpha-tubulin gene family at the levelof development and gene regulation and also at the level of the genome.

MATERIALS AND METHODS

cDNA library screening.—The cDNA library used in this study was a giftfrom Jo Ann Banks (Purdue University). It was made from 12-day-old culturesof C. richardii gametophytes of the Hn-n strain containing both males andhermaphrodites. The cDNA was cloned into the EcoRI site of the lambdaZipLox bacteriophage vector (Life Technologies). The aliquot used to screenfor alpha-tubulin genes was from a sample that had been amplified from theoriginal library. The library was screened using standard methods for detectingspecific proteins via antibody labeling (Young and Davis, 1991). In brief, thebacteriophage vector was first grown on a lawn of bacteria and, after plaqueswere produced, gene expression was induced with IPTG. Proteins from thelibrary were adhered to nylon membranes and detected by hybridization to ananti alpha-tubulin monoclonal antibody (Sigma, catalogue number T5168).The presence of this antibody was detected with a labeled secondary antibody.After initial detection and isolation of plaques that tested positive forexpression of alpha-tubulin, several rounds of re-screening were conducted

SCOTT ET AL.: ALPHA-TUBULIN GENE FAMILY IN CERATOPTERIS RICHARDII 49

to eliminate contaminating vectors that were not positive for alpha-tubulin.Bacteriophage vectors containing alpha-tubulin cDNAs were converted toplasmids via an in vivo excision process as described in technical referencematerials supplied by Life Technologies.

DNA sequencing.—Plasmids with the cDNA inserts were used as thesequencing templates. For each plasmid, at least two sequencing reactionswere conducted using the two primer sites that flank the EcoRI site where thecDNA is inserted. When the results from two sequencing reactions indicatedthat the insert was longer than 1220 base pairs, two additional pairs of primerswere constructed based on the new sequence data. These new internalsequencing primers provided more reliable results for the central regions ofthese longer inserts. The sequence reads from either two or four reactions wereassembled using the software program ContigExpress (NTI Vector). Ambigu-ities that resulted from discrepancies between overlapping sequences wereresolved by relying on the sequence(s) with the clearest chromatogram.

BLAST comparisons.—The sequences were compared using the standardBLAST program for proteins (protein-protein BLAST) provided on the NCBIweb site to compare individual sequences with other sequences contained inthe GenBank database as described in the Results section. The Composition-based statistic option and the filter options were not enabled so that the resultsof the BLAST comparisons would be based solely on similarities and/ordifferences of individual amino acids. Since the GenBank database isconstantly being updated, it should be noted that the reported BLASTcomparisons were last confirmed on July 29, 2005.

Phylogenetic analysis of alpha-tubulin gene relationships.—A BayesianInference analysis of nucleotide sequences was used to infer the phylogeneticrelationships of the four C. richardii alpha-tubulin genes with regard to otherplant alpha-tubulin gene nucleotide sequences available in GenBank forViridiplantae. Sequences used were from a broad selection of plants, includingthose corresponding to the results of the amino acid BLAST search above,except that nucleotide sequences precisely corresponding to the amino acidsequences of Gossypium hirsutum Q6VAG0 and Chlamydomonas reinhardtiiP09204 could not be located in GenBank. Prior to analysis, nucleotidesequences were aligned visually. The first three nucleotides, which wereinvariable, and the last 45 nucleotides, which caused ambiguous alignments,were removed. MrModeltest 2.2, a modified version of Modeltest 3.6 (Posadaand Crandall, 1998) determined that the suitable nucleotide substitutionmodel for the analyzed sequences is GTR+I+G. Bayesian analysis wasconducted using MrBayes 3.1 (Huelsenbeck and Ronquist, 2001; Ronquistand Huelsenbeck, 2003) with three million generations and a samplefrequency of 1000. The first 300 ‘‘burn-in’’ trees were discarded after theanalysis. Five independent runs using the same setting converged to identicaltrees, except that one node was resolved in only one run. Two beta-tubulingenes were used as an outgroup to root the resulting tree. The species used andtheir GenBank accession numbers are identified in the Results section, as arethe Bayesian posterior probability confidence values for the tree’s clades.


Linkage mapping of the alpha-tubulin genes.—A genetic linkage map of theCeratopteris richardii genome was developed independently of this alpha-tubulin study (Nakazato et al., 2006). The mapping population of ,500Doubled Haploid Lines (DHLs) was generated by intragametophytic selfing ofgametophytes derived from spores of an initial cross between diploid inbredlines of highly diverged geographic races of C. richardii (Hickok et al., 1995):WN8 (derived from a Nicaraguan collection, Nichols 1719, GH) and Ha-PQ45,a mutant of Hn-n (derived from a Cuban collection, Killip 44595, GH). The mapis based on analysis of Restriction Fragment Length Polymorphism (RFLP),Amplified Fragment Length Polymorphism (AFLP), and allozyme markers.RFLPs used in the general mapping project were generated from genomic DNAof parental and DHL sporophytes digested with EcoRI and HindIII, separatedon 0.8% agarose gels in 1X TAE, and Southern blotted to nylon membranes.Further details of the methods used in development of mapping materials andmap construction are in Nakazato and Gastony (2006) and Nakazato et al.(2006). Alpha-tubulin gene copies were located on the linkage map in thefollowing way. RFLPs containing the alpha-tubulin genes were detected byprobing the Southern blots with an alpha-tubulin probe made cheiluminescentby digoxigenin (DIG)-labeling according to a protocol optimized for the C.richardii genome mapping project. The alpha-tubulin probe used is C.richardii cDNA clone Cri_10_E18_SP6 (GenBank sequence num-ber BQ086953), from the cDNA library of C. richardii gametophytic tissueprovided by Jo Ann Banks (Purdue University). This clone sequencecorresponds to GenBank sequence AY231146 of TuaCR1 in Table 1, accordingto an NCBI BLAST search using blastn, which found sequence identity at 651/652 (99.85%) of the bases compared. Probing of the mapping project’sSouthern blots with the alpha-tubulin probe was carried out toward the endof the mapping project when the Southern blots were beginning to wear out.Thus the quality of the autorads presented here are suboptimal but neverthe-less scorable. Parental alpha-tubulin RFLPs segregating in the DHL mappingpopulation were scored and placed on the linkage maps by MAPMAKER/EXP3.0 (Lander et al., 1987) at a UNIX workstation at Indiana University,Department of Biology, at settings used for the general mapping project.

RESULTS

Sequencing results.—A cursory comparison of the amino acid sequences forthe four Ceratopteris alpha-tubulin genes (Fig. 1) suggests immediately thatthey are relatively similar to one another. Although only one of the foursequences is complete, it is possible to note at least two trends directly fromthe comparison presented (Fig. 1). First, for the portion where sequences areavailable for all four genes (i.e., residues 195 to 332 of TuaCR1), TuaCR3 andTuaCR4 are the most distinct. In this region, TuaCR3 contains five uniqueamino acid residues and TuaCR4 contains four, while the other two sequenceseach possess only a single unique amino acid. The second observation that canbe made directly from these data relates only to the three sequences that


include complete carboxy-terminal ends (i.e., TuaCR1, 2, and 3). In this region,these three sequences are highly diverse, with TuaCR3 once again exhibitingthe greatest number of unique residues.

An additional sequence identified as an alpha-tubulin sequence fromCeratopteris has been deposited in GenBank by a separate research group(Salmi et al., 2005). This 825 base-long sequence (GenBank accessionBE642799) is a single pass sequence from a collection of expressed sequencetags. The bl2seq tool on the NCBI web site was used to compare this sequencewith the four alpha-tubulin sequences described here (data not shown). A largeregion of this sequence (ranging from 574 to 640 bases) is 77 or 78% identicalto the beginning portions of TuaCR1 and TuaCR4 respectively. It has nosignificant similarity to TuaCR3. The highest identity (88% for a region that is195 bases long) occurs between this sequence and TuaCR2. This region ofsimilarity occurs at the beginning of TuaCR2 and at the end of the sequencedescribed by Salmi et al. (2005). Because this sequence was derived froma single sequencing experiment, it may be expected to contain a relatively largenumber of incorrect bases. Furthermore, the authors did not provide a deduced

TABLE 1. A comparison of the ten Ceratopteris alpha-tubulin sequences identified in this study to

other alpha-tubulin sequences recorded in GenBank.

Names for

new gene

GenBank

Accession

number

Number of bases

sequenced in

DNA1

Size of

Predicted

ORF

Accession number, description, and

identity for amino acid sequence(s)2 most

similar to new sequences3

TuaCR1 AY231146 1,676 450 1) Q6VAG0, Tubulin alpha-2 chain

AY862561 1,159 306 Gossypium hirsutum, 436/450 (96.2%)

AY862563 1,564 450

TuaCR2 AY862565 1,015 266 1) AAK81858.1, alpha tubulin subunit

AY862566 1,016 267 Rosa Hybrid cultivar, 260/267 (97.4%)

AY862567 1,018 267 2) AAV92379.1, alpha tubulin 1

AY862569 1,016 267 Pseudotsuga menziesii, 260/267 (97.4%)

TuaCR3 AY862562 938 257 1) BAA03955, alpha-tubulin 1

AY862564 1044 257 Chlorella vulgaris 245/257 (95.3%)

2) P09204, Tubulin alpha-1 chain

Chlamydomonas reinhardtii, 245/257

(95.3%)

TuaCR4 AY862568 975 325 1) AAV92379.1, alpha tubulin 1

Pseudotsuga menziesii, 316/325 (97.2%)

1This includes reliable DNA sequences before and after the open reading frame if present, but not

the predicted polyA tail.2If two or more versions of the same gene occurred in GenBank, only the most recent one is listed

here. If several identical sequences were submitted by the same author(s) and are listed with the

same submission date, only the one with the last accession number in the numerical sequence is

listed here. The descriptions are given in the form in which they were submitted to GenBank.3Sequence similarities were determined using the standard BLAST program for protein sequences

to search the GenBank database. The Composition-based statistic option and the filter options were

not enabled. The BLAST searches were performed on July 29, 2005.


FIG. 1. Alignment of the predicted amino acid sequences of four alpha-tubulin proteins from

Ceratopteris richardii; the standard symbols for amino acids are used. Only the sequence for

TuaCR1 is complete, and all other proteins are compared to it. Dashes represent unknown amino

acids, asterisks represent known amino acids that are identical to those shown for TuaCR1, bold

letters represent known amino acids that differ between TuaCR1 and at least one other TuaCR

copy, and a blank space represents an apparent gap in the alignment.


protein sequence for this gene. For these reasons, and because it may representan otherwise uncharacterized portion of one of the genes described here, it wasnot included in any of the comparisons described below.

A BLAST comparison of the four protein sequences for the Ceratopterisgenes with other alpha-tubulin sequences recorded in GenBank revealsa somewhat diverse pattern (see the last column of Table 1). First, each ofthe four sequences was more similar to an alpha-tubulin from a different plantspecies than it was to any other Ceratopteris alpha-tubulin sequence. Second,the sequences from other plants that are most similar to the Ceratopterissequences are quite diverse. The first two Ceratopteris amino acid sequences(those of TuaCR1 and TuaCR2) are most similar to sequences derived fromdicotyledonous plants and/or a gymnosperm, the third is most similar to twodifferent algal sequences, and the fourth is most similar to a sequence froma gymnosperm. These observations are perhaps more noteworthy in light of thefact that two distinct sequences for alpha-tubulin genes from the fern Anemiaphyllitidis (one is a partial sequence) and two from the moss Physcomitrellapatens are deposited in GenBank. When compared to the Ceratopteris genes(data not shown), some of the four lower plant genes have nearly as manyidentical amino acid sequences as do the sequences from the higher plants andalgae shown in Table 1. However, when these high levels of identity exist, thelower plant sequences also have at least one additional amino acid or at leastone less amino acid than the Ceratopteris sequences, resulting in gaps in thesequence homologies. The homologies indicated in Table 1 do not have suchgaps and therefore the sequences they refer to are considered to be moresimilar to the Ceratopteris sequences.

For comparative purposes, BLAST searches were conducted for each of theknown alpha-tubulin genes of Arabidopsis to determine what sequences fromArabidopsis or other organisms would show the highest similarities to each ofthese genes (data not shown). The entire genome of Arabidopsis has beensequenced, and six separate alpha-tubulin genes (TUA1–TUA6) have beenlocated in its genome. The patterns of similarity for these genes aredramatically different from those of Ceratopteris. In general, the sixArabidopsis alpha-tubulin genes show much more similarity to one anotherthan do the four Ceratopteris genes. Among these six genes are two pairswhose protein sequences are identical. The proteins of TUA2 and TUA4 areidentical, and those of TUA3 and TUA5 are identical. Furthermore, TUA6 isalmost identical to TUA2/4, sharing 448 out of 450 amino acids (99.6%identity). The amino acid sequences that are the next most similar to these fiveArabidopsis genes (after comparing them to other Arabidopsis genes) are allsequences from angiosperms. The sequence of TUA2/4 is 98.7% identical toa sequence from Brassica napus, the sequence of TUA6 is 99.1% identical tothe same sequence from B. napus, and the sequence of TUA3/5 is 97.1%identical to a sequence from Oryza sativa. Only the sequence of the remainingArabidopsis gene product, that of TUA1, is more similar to the sequence ofa different plant than it is to another Arabidopsis gene. The TUA1 proteinshares 414 amino acids out of 450 (92% identity) with the grass Miscanthus.


To further compare the relative diversity of the Ceratopteris and Arabidopsisalpha-tubulin proteins, the sequences from the carboxy terminus region,which is considered to be highly variable in alpha-tubulin proteins (e.g., seeSullivan, 1988; Fosket and Morejohn, 1992), were aligned with one another(Fig. 2). Since only three of the four Ceratopteris sequences contain the codingregion for the carboxy terminus portion of the protein, TuaCR4 was omittedfrom this analysis. For both species the amino acid sequence is relativelyconserved until the last 14 residues at the carboxy terminus end, or in the caseof TuaCR1, which appears to have a single amino acid deletion in this region,the last 13 amino acids. In this region the Arabidopsis sequences share 6identical amino acids out of 14, while the Ceratopteris sequences share only 3identical amino acids.

Gene phylogeny results.—The Bayesian tree (Fig. 3) infers phylogeneticrelationships of nucleotide sequences of the four sequenced C. richardii alpha-tubulin genes in relation to alpha-tubulin genes from other plant species. Thisfigure depicts the tree from the single run noted in Materials and Methods asresolving one of the nodes, the resolved subnodes being toward the top of thetree with probabilities of 0.80, 0.50, and 0.81. The central clade from Zea mays22149 to Oryza sativa japonica cultivargroup 1136121 was not resolved in anyof the Bayesian runs. Posterior probability values show that many clades ofthis gene tree are strongly to moderately supported, although some deeperbranches had weak support. The unresolved and weakly supported clades,however, are inconsequential to this paper, which focuses instead on the fourcopies of C. richardii genes and their placement in strongly supported clades.All nodes separating the four Ceratopteris genes are strongly supported. ThreeCeratopteris copies (TuaCR1, TuaCR2, and TuaCR4) toward the top of Fig. 3are relatively closely related to each other in a strongly supported subcladethat is part of the major clade stretching from C. richardii TuaCR1 29423812 to

FIG. 2. Comparison of the carboxy terminus regions of Arabidopsis and Ceratopteris alpha-tubulin

genes. Bold letters represent regions where the sequences within each set are not identical; a blank

space represents an apparent gap in the alignment.


FIG. 3. Bayesian inference gene tree based on aligned nucleotide sequences of alpha-tubulin genes

of Viridiplantae in GenBank (excluding the invariant first three nucleotides and the non-alignable

last 45 nucleotides), showing relationships of the four sequenced C. richardii alpha-tubulin gene

copies discussed. Taxon sources and GenBank gi numbers are given for each gene copy. Bayesian

posterior probability values indicate support for clades.


Oryza sativa japonica cultivargroup 1136121. This major clade is separatedfrom its sister clade of miscellaneous algae and C. richardii TuaCR3 57867856with a posterior probability of 1.00, although the precise positioning of C.richardii TuaCR3 in the algal clade is less strongly supported.

Mapping results.—Probings of parental and DHL DNAs with the DIG-labeledalpha-tubulin Cri_10_E18_SP6 cDNA clone yielded at least ca. 13–17restriction fragment bands per DHL. The total number of bands cannot bedetermined with precision because of overlap and faintness of some of thebands. As an example, Fig. 4 shows the probing results for parentalsporophytes Ha-PQ45 and WN8 and for ten DHLs whose genomic DNAs werecut with restriction enzymes HindIII and EcoRI, respectively. The DHL in lane2 of the HindIII digest, for example, probably exhibits at least 13 bands(arrowheads) containing sequence to which the alpha-tubulin probe annealsand the DHL in lane 10 shows at least 17 bands. In some cases multiple bandsmay represent a single alpha-tubulin gene sequence that has been cut by theHindIII or EcoRI restriction enzyme. Although the sequence of the cDNA probeused in this alpha-tubulin study contains no HindIII or EcoRI restriction site,the coding region of the target DNA might contain such sites, and intronswithin the genes of the target DNA may contain HindIII or EcoRI sites. No dataare available to address these possibilities.

Mapping can be performed only for those parental alpha-tubulin gene copiescontained in restriction fragments that are polymorphic between the twoparents and that therefore segregate in the DHLs. Four such unequivocalsegregating sets of restriction fragments were observed in the genomic DNAsdigested with HindIII and EcoRI in our mapping population. These areidentified (Fig. 4) as H20046154_1 and H20046154_2 scored from the HindIII

FIG. 4. Example autoradiograms of total genomic DNAs from the two parental C. richardii races

and 10 DHLs derived from their cross, all probed with a DIG-labeled alpha-tubulin Cri_10_E18_SP6

cDNA clone after respective digestions of the genomic DNAs with HindIII or EcoRI. Arrowheads in

DHL lanes 2 and 10 indicate at least 13 and 17 probable bands respectively. Segregating co-

dominant and dominant RFLP bands containing alpha-tubulin genes are identified by arrows to the

left of each autoradiogram and are discussed in the text and mapped in Fig. 5.


digest and E20046154_1 E20046154_2 scored from the EcoRI digest. Poly-morphic restriction fragment markers are usually expressed as co-dominantbands, meaning that the gene’s presence is visualized in respective bands fromboth parents. In Fig. 4, this is exemplified in the HindIII digest by locusH20046154_1 where the larger fragment from parent Ha-PQ45 (also seen inDHL lanes 1, 3, 6, 10) is ca. 1 mm closer to the top of the figure than is thesmaller fragment from parent WN8 (also seen in DHL lanes 2, 4, 5, 7, 8, 9).Locus E20046154_1 in the EcoRI digest in Fig. 4 shows a similar pattern. Inthis case the smaller fragment in parent Ha-PQ45 (also seen in DHL lanes 1, 3,6, 10) is ca. 1.5 mm farther from the top of the autorad than is the largerfragment from parent WN8 (also seen in DHL lanes 2, 4, 5, 7, 8, 9). The twobands from the HindIII digest co-segregate with the two bands from the EcoRIdigest in all of the mapping population’s DHLs, indicating that they eithermark the same identical locus visualized in the two different digests or marktwo loci so closely linked that they show no crossover distance between them(i.e., they map to the same location). Locus H20046154_2 on the HindIII digest,on the other hand, is visualized as a dominant band in parent WN8 (also seenin DHL lanes 2, 4, 5, 6, 7, 8) meaning that no alternative band expression isvisualized in the Ha-PQ45 parent (and in DHL lanes 1, 3, 9, 10). The reason forthis dominant pattern is presently unknown but may be because theH20046154_2 gene copy in the Ha-PQ45 parent has been lost or has beenmoved to a different part of the genome where it cannot be scored because itoverlaps with a different band, etc. In the EcoRI digest, locus E20046154_2 isalso expressed as a dominant marker, present in parent Ha-PQ45 (and in DHLlanes 1, 4, 5, 9, 10), but with an alternative band lacking in parent WN8 (and inDHL lanes 2, 3, 6, 7, 8). Clearly the bands marking H20046154_2 andE20046154_2 do not co-segregate in the DHLs of the two digests, indicatingthat they mark different loci. The mapping program places H20046154_2 onlinkage group (LG) 28 and E20046154_2 on LG 24 (Fig. 5 at arrows), whereasH20046154_1 and E20046154_1 map to the same position on LG 17 (Fig. 5arrow).

DISCUSSION

Map positions of alpha-tubulin genes.—The genetic linkage mapping projectfor C. richardii, fully described by Nakazato et al. (2006), has identified 41linkage groups that partly correspond to the 39 chromosomes per haploid setin this species. Alpha-tubulin gene copy H20046154_2 is located on LG 28,H&E20046154_1 is on LG 17, and E20046154_2 is on LG 24 (Fig. 5). Also seenon these three linkage groups are a large number of AFLP markers, eachidentified by a lowercase ‘‘a’’ followed by the primer pair used to generate it,and additional RFLP markers whose code numbers are the GenBank ginumbers of the cDNA library clones used as probes, prefaced with an ‘‘H’’ oran ‘‘E’’ to indicate whether that locus was visualized on the HindIII or theEcoRI digests. On LG 17 is also found the isocitrate dehydrogenase (IDH)isozyme locus, which maps to virtually the same position as H9960276_2. In


all cases, the cumulative distance in centiMorgans from the topmost maker isgiven to the left of each linkage group.

Because the alpha-tubulin clone used as a DIG-labeled probe here anneals toevery alpha-tubulin gene copy with sufficient sequence identity, given thestringency conditions of our general mapping project, we cannot determinewhich of the genes in Table 1 are represented by respective markersH&E20046154_1, H20046154_2, and E20046154_2. The large number of bands(at least 13 to 17 in the indicated lanes in Fig. 4) suggests a large alpha-tubulingene family in C. richardii, even if some genes are represented by more thanone band. We were unable to map more than three or four alpha-tubulin locibecause only four loci were contained in restriction fragments polymorphic inthe two parents in HindIII and EcoRI digests.

Ceratopteris alpha-tubulin gene phylogeny.—The phylogenetic relation-ships of the four C. richardii genes in this paper are illustrated in Fig. 3.Bayesian posterior probability values show that copies TuaCR1 and TuaCR2 atthe top of the tree are very strongly grouped as sister to each other and thatboth are in turn robustly sister to a copy from the conifer Pseudotsugamenziesii. Also with very strong support, the preceding are separated fromTuaCR4 which is sister to an alpha-tubulin gene copy from the fern Anemiaphyllitidis. The most parsimonious hypothesis to explain these present data isthat TuaCR1 and TuaCR2 result from a recent duplication, perhaps after thedivergence of ferns and seed plants, and that the gene duplication leading tothe TuaCR4 lineage and the TuaCR1+TuaCR2+Pseudotsuga lineage precededthe divergence of ferns and seed plants. Ceratopteris richardii TuaCR3 (Fig. 3),on the other hand, is grouped in a clade of algal alpha-tubulin gene copiesbelow the center of the tree. Although the precise relationships within this

FIG. 5. Three of the 41 linkage groups presently identified in the C. richardii genome, showing at

arrows the positions of the four segregating RFLP bands that contain alpha-tubulin genes identified

in Fig. 4.


clade of algal genes+TuaCR3 are not strongly supported, a 1.0 posteriorprobability value strongly separates this clade from the other three Ceratopterisgenes in the major sister clade of miscellaneous seed plant alpha-tubulin genecopies above it. This indicates that TuaCR3 had already diverged from thecommon ancestor of the other three Ceratopteris alpha-tubulin genes by thetime of the common ancestor of the ferns and algae, preceding the divergenceof algae and land plants. The sampling of alpha-tubulin genes from plants ingeneral, however, is currently very incomplete. The four C. richardii copiesdiscussed here are simply those expressed in a cDNA library derived from 12-day-old gametophytes. Surely they do not represent the full range of alpha-tubulin gene copies expressed in the life cycle of C. richardii, just as the twoalpha-tubulin gene copies sequenced thus far from the fern Anemia phyllitidisin Fig. 3 must represent only a fraction of the copies from that species. Inferredtimings of duplications will likely change when more alpha-tubulin genesequences become available. It may be noteworthy that C. richardii and thealgae have motile gametes whereas all of the seed plants in Fig. 3 (includingPseudotsuga) lack motile sperm. This suggests TuaCR3 as a candidate for thegene copy functioning in sperm motility microtubules.

The occurrence of distantly related beta- and gamma-tubulin genes resultingfrom ancient duplications and the large number of alpha-tubulin gene copiesseen in the C. richardii genomes in Fig. 4 indicate that there has been extensiveand continuous duplication of tubulin genes throughout the evolution oforganisms. It is also likely that deletion/silencing of tubulin gene copies isfrequent, perhaps in lineage-specific ways. For example, despite intensivesequencing of alpha-tubulin genes in seed plants, gene copies from seed plantsare absent from the lineage containing the TuaCR3 and algal gene copies inFig. 3. This suggests that silencing/deletion of this copy may have occurredspecifically in the seed plant lineage.

Diversity of the C. richardii alpha-tubulin genes.—In terms of amino acidsequences, particularly the carboxy termini, the four Ceratopteris alpha-tubulin sequences appear to be relatively diverse both when compared to oneanother (Figs. 1 and 2) and when compared to sequences from other plants (seeTable 1, Fig. 2, and text of the Results section). The structural similarity of theCeratopteris alpha-tubulin protein sequences to those of other diverse speciesis not surprising. This kind of similarity is often seen for alpha-tubulinproteins when sequences are compared using the BLAST algorithm (data notshown). The phylogenetic relationships of these four C. richardii sequencesdetermined by Bayesian analysis of their nucleotide sequences excluding theinvariant first codon and the non alignable 45 carboxy terminus nucleotides,however, indicates less diversity (Fig. 3). Phylogenetic analysis shows thatTuaCR1 is closest to TuaCR2 and that, except for the alpha-tubulin copies fromPseudotsuga and Anemia, TuaCR4 is most closely related to TuaCR1 andTuaCR2. TuaCR3, on the other hand, is quite distant from the other three C.richardii copies and is most closely related to algal alpha-tubulin genes.

Potential insights from further studies of the Ceratopteris alpha-tubulingene family.—Two main types of insights may be gained by studying the


diverse genes of organisms like Ceratopteris, and these are discussed in turnbelow. First, the diversity maintained in gene families like that of alpha-tubulin may provide new insights into the evolutionary history of specificgenes as well as the mechanisms by which they evolve. The conventionalwisdom regarding gene families is that the multiple versions of related genesdevelop by gene duplication and random mutation that produces variantssubject to varying degrees of selection. In some cases, the different forms ofthese related genes assume different roles over the course of evolutionaryhistory (e.g., see Zhang, 2003; Irish and Litt, 2005).

Consider the comparison presented earlier between the alpha-tubulin genesof Arabidopsis and those of Ceratopteris. The relative similarity in amino acidsequence of the six alpha-tubulin genes of Arabidopsis thaliana noted abovesuggests at least two likely interpretations. First, selection pressure may bevery high for alpha-tubulin genes in this species, such that very little deviationis tolerated. Second, there may be relatively fewer potential roles for alpha-tubulin in the development and life history of Arabidopsis (for example, spermare non-motile, unlike the situation in ferns) and therefore evolutionary factorshave resulted in less diversity. In this respect, it is noteworthy that theArabidopsis alpha-tubulin gene that differs most from the other five, TUA1,appears to be preferentially expressed primarily in pollen grains (Carpenter etal. 1992), while the others are apparently expressed in various tissuesthroughout the plant (Kopczak et al., 1992).

By comparison with the alpha-tubulin genes of Arabidopsis, amino acidsequences of the four known alpha-tubulin genes of Ceratopteris appearsomewhat diverse, suggesting selective pressures favoring the origin ormaintenance of alpha-tubulin variation in this fern. This may relate in partto its fully independent gametophyte generation with perhaps more potentialroles for alpha-tubulin (for example, as a component of the flagella inswimming sperm) than are found in its angiosperm counterpart, Arabidopsis.If these considerations prove correct and generally applicable to other genes,new insights regarding evolutionary mechanisms could be gained from studiesfocusing on C. richardii as a model homosporous vascular plant that couldnever be gained by studying plants like Arabidopsis alone.

A second type of information to be gained by studying diverse members ofgene families, such as those of the C. richardii alpha-tubulin family, relates togene expression. For example, what are the specific roles of the variousmembers of the gene family? When during development and under whichenvironmental conditions are these genes expressed? How is the expression ofeach gene controlled? To provide some insight into how C. richardii can beutilized to answer such questions, two potential research strategies utilizingthe new sequence data presented in this report are described below.

A large number of morphological mutants exist from classical mutagenesisand selection screens using Ceratopteris (reviewed in Hickok, 1987; Hickok etal., 1987, 1995). Some affect cell shape, like the dwarf or bubbles mutant(Hickok, personal communication), others affect intracellular organization,like the polka-dot mutant (Vaughn et al., 1990), while yet others, like sleepy


sperm (Renzaglia et al., 2004) affect the mobility of the sperm. Based on thephenotypes of the above mutants, it is likely that the defect associated withsome of them may involve either an alpha-tubulin gene itself, a gene thatcontrols expression of an alpha-tubulin gene, or a gene for another protein thatinteracts with alpha-tubulin. With the four sequences presented in this report,it will be relatively simple to assess the first two possibilities, using PCR toobtain sequence data and techniques like reverse transcription-PCR (RT-PCR)and/or real-time PCR to study relative levels of gene expression.

One may also study the roles of the various alpha-tubulin genes by utilizingrecently developed methods for inducing gene silencing via RNA interference(RNAi) in spores and gametophytes of Ceratopteris (Stout et al., 2003;Rutherford et al., 2004). RNAi is a technique whereby specific genes may besilenced by the introduction of short single- or double-stranded RNAsequences into an organism’s cells. These short RNA sequences, which mustbe highly similar to or identical to the complementary coding sequenceswithin the targeted gene, trigger a cellular response that causes the degradationof the mRNA transcribed from the targeted gene.

In the two studies published so far using RNAi in C. richardii, the shortRNAs have been introduced either by simply soaking spores directly ina solution containing the RNAs (Stout et al., 2003) or by using the biolisticmethod to deliver the RNA into gametophytic cells (Rutherford et al., 2004).The effects of gene silencing in C. richardii appear to be transmissible to cellsdescended from the cell into which the inhibiting RNA was initiallyintroduced. In experiments where the inhibitory RNA was introduced intocells of young gametophytes by biolistic delivery, an observable phenotypewas sometimes even transmitted to developing sporophytes (Rutherford et al.,2004).

Each of the four alpha-tubulin genes described here has at least severalportions of its sequence that are not shared by the other known members of thisgene family (this is especially evident in the 3’ regions of TuaCR1, 2, and 3).These sequences can be used to design short inhibitory RNAs that should beable to silence specifically the expression of their respective genes.Alternatively, it should be possible to silence two or more of these genes atthe same time by using short RNA sequences that are shared by two or moregenes. In either approach, such silencing can be confirmed by demonstratingan absence of (or reduction in) the targeted mRNA by using RT-PCR (Stout etal., 2003) or real time PCR (Rutherford et al., 2004), and phenotypes resultingfrom the silencing may also be observed in some cases (Rutherford et al., 2004).

Several potential phenotypes could be easily observable as a result ofsilencing alpha-tubulin expression in Ceratopteris. Some may include suchdevelopmental defects as the inhibition of spore germination, or the generationof abnormal rhizoids, prothalial cells, or specialized cells such as those ofantheridia or archegonia. Cytoskeletal defects generated by this approach mayalso lead to observable phenotypes—some may be associated with cell size orshape, while others may be as striking as the phenotype of the polka-dotmutant. Other observable abnormalities may affect the function of spermata-


zoids, generating phenotypes similar to the sleepy sperm mutant. The obviousadvantage to generating such phenotypes through RNAi is that in each case,the phenotype will be directly associated with the silencing of one or morespecific alpha-tubulin genes. This in turn may lead to the assignment ofspecific roles for each member of this gene family.

The linkage positions of the three mapped loci together with the specificsequencing data provided here should facilitate future cloning of genomicsequences containing alpha-tubulin genes in C. richardii. This will enablecharacterization of the organization and controlling elements of these genes,enhancing our understanding of the evolution, function, and regulation of thealpha-tubulin gene family in vascular plants in general. The researchpossibilities discussed here illustrate how robust molecular techniques canbe applied to the C. richardii system, furthering its usefulness as a modelsystem.

ACKNOWLEDGMENTS

RJS and JWW thank Nalco Chemical Company for funds used by Wheaton College to purchase

a DNA sequencer, and the Wheaton College Administration and the Wheaton College Alumni

Association for funds used to conduct the sequencing work. GJG and TN thank the National

Science Foundation for support of their mapping project through grant DEB-0128926.

LITERATURE CITED

ASO, K., M. KATO, J. A. BANKS and M. HASEBE. 1999. Characterization of homeodomain-leucine

zipper genes in the fern Ceratopteris richardii and the evolution of the homeodomain-leucine

zipper gene family in vascular plants. Molec. Biol. Evol. 16:544–552.

BANKS, J. A. 1999. Gametophyte development in ferns. Annual Rev. Pl. Physiol. Pl. Molec. Biol.

50:163–186.

CARPENTER, J. L., S. E. PLOENSE, D. P. SNUSTAD and D. SILFLOW. 1992. Preferential expression of an

a-tubulin gene of Arabidopsis in pollen. Pl. Cell 4:557–571.

FOSKET, D. E. and L. C. MOREJOHN. 1992. Structural and functional organization of tubulin. Annual

Rev. Pl. Physiol. Pl. Molec. Biol. 43:201–240.

HASEBE, M., C-K. WEN and J. A. BANKS. 1998. Characterization of MADS homeotic genes in the fern

Ceratopteris richardii. Proc. Natl. Acad. Sci. U.S.A. 95:6222–6227.

HICKOK, L. G. 1976. Cytological relationships between three diploid species of the fern genus

Ceratopteris. Canad. J. Bot. 55:1660–1667.

HICKOK, L. G. 1977a. An apomictic mutant for sticky chromosomes in the fern Ceratopteris. Canad.

J. Bot. 55:2186–2195.

HICKOK, L. G. 1977b. The cytology and derivation of a temperature-sensitive meiotic mutant in the

fern Ceratopteris. Amer. J. Bot. 64:552–563.

HICKOK, L. G. 1978. Homoeologous chromosome pairing and restricted segregation in the fern

Ceratopteris. Amer. J. Bot. 65:516–521.

HICKOK, L. G. 1979a. Apogamy and somatic restitution in the fern Ceratopteris. Amer. J. Bot.

66:1074–1078.

HICKOK, L. G. 1979b. A cytological study of intraspecific variation in Ceratopteris thalictroides.

Canad. J. Bot. 57:1694–1700.

HICKOK, L. G. 1987. Applications of in vitro selection systems: whole plant selection using the

haploid phase of the fern Ceratopteris, pp. 53–56. In H. M. LeBaron, R. O. Muma, R. C.

Honeycutt and J. H. Duesing, eds. Biotechnology in agricultural chemistry. American

Chemical Society, Washington, DC.


HICKOK, L. G. and E. J. KLEKOWSKI. 1973. Abnormal reductional and nonreductional meiosis in

Ceratopteris: Alternatives to homozygosity and hybrid sterility in homosporous ferns. Amer.

J. Bot. 60:1010–1022.

HICKOK, L. G. and E. J. KLEKOWSKI. 1974. Inchoate speciation in Ceratopteris: An analysis of the

synthesized hybrid C. richardii 3 C. pteridoides. Evolution 28:439–446.

HICKOK, L. G., T. R. WARNE and R. S. FRIBOURG. 1995. The biology of the fern Ceratopteris and its use

as a model system. Int. J. Pl. Sci. 156:332–345.

HICKOK, L. G., T. R. WARNE and M. K. SLOCUM. 1987. Ceratopteris richardii: Applications for

experimental plant biology. Amer. J. Bot. 74:1304–1316.

HOFFMAN, J. C. and K. C. VAUGHN. 1995a. Using the developing spermatogenous cells of Ceratopteristo unlock the mysteries of the plant cytoskeleton. Int. J. Pl. Sci. 156:346–358.

HOFFMAN, J. C. and K. C. VAUGHN. 1995b. Post-translational tubulin modifications in spermatoge-

neous cells of the pteridophyte Ceratopteris richardii. Protoplasma 186:169–182.

HOFFMAN, J. C. and K. C. VAUGHN. 1996. Spline and flagellar microtubules are resistant to mitotic

disrupter herbicides. Protoplasma 192:57–69.

HOFFMAN, J. C., K. C. VAUGHN and H. C. JOSHI. 1994. Structural and immunocytochemical

characterization of microtubule organizing centers in pteridophyte spermatogenous cells.

Protoplasma 179:46–60.

HUELSENBECK, J. P. and F. RONQUIST. 2001. MRBAYES: Bayesian inference of phylogenetic trees.

Bioinformatics 17:754–755.

IRISH, V. F. and A. LITT. 2005. Flower development and evolution: gene duplication, diversification

and redeployment. Current Opinion in Genetics & Development 15:454–460.

KOPCZAK, S. D., N. A. HAAS, P. J. HUSSEY, C. D. SILFLOW and D. P. SNUSTAD. 1992. The small genome of

Arabidopsis contains at least six expressed a-tubulin genes. Pl. Cell 4:539–547.

LANDER, E., J. ABRAHAMSON, A. BARLOW, M. DALY, S. LINCOLN, L. NEWBURG and P. GREEN. 1987.

MAPMAKER: A computer package for constructing genetic linkage maps. Cytogenet. & Cell

Genet. 46:642.

MUNSTER, T., J. PAHNKE, A. DI ROSA, J. T. KIM, W. MARTIN, H. SAEDLER and G. THEISSEN. 1997. Floral

homeotic genes were recruited from homologous MADS-box genes preexisting in the

common ancestor of ferns and seed plants. Proc. Natl. Acad. Sci. U.S.A. 94:2415–2420.

MURATA, T., A. KADOTA and M. WADA. 1997. Effects of blue light on cell elongation and microtuble

orientation in dark-grown gametophytes of Ceratopteris richardii. Pl. Cell Physiol. 38:201–209.

NAKAZATO, T. and G. J. GASTONY. 2006. High-throughput RFLP genotyping method for large genomes

based on a chemiluminescent detection system. Plant Molec. Biol. Rep. 24:245a–245f.

NAKAZATO, T., M.-K. JUNG, E. A. HOUSWORTH, L. H. RIESEBERG and G. J. GASTONY. 2006. Genetic map-

based analysis of genome structure in the homosporous fern Ceratopteris richardii. Genetics

173:1585–1597.

POSADA, D. and K. A. CRANDALL. 1998. MODELTEST: testing the model of DNA substitution.

Bioinformatics 14:817–818.

PRYER, K. M., H. SCHNEIDER, A. R. SMITH, R. CRANFILL, P. G. WOLF, J. S. HUNT and S. D. SIPES. 2001.

Horsetails and ferns are a monophyletic group and the closest living relatives to seed plants.

Nature 409:618–622.

RAGHAVAN, V. 1989. Developmental Biology of Fern Gametophytes. Cambridge University Press,

New York, NY.

RENZAGLIA, K. S., K. D. WOOD, G. RUPP and L. G. HICKOK. 2004. Characterization of the sleepy sperm

mutant in the fern Ceratopteris richardii: A new model for the study of axonemal function.

Canad. J. Bot. 82:1602–1617.

RONQUIST, F. and J. P. HUELSENBECK. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed

models. Bioinformatics 19:1572–1574.

RUTHERFORD, G., M. TANURDZIC, M. HASEBE and J. A. BANKS. 2004. A systemic gene silencing method

suitable for high throughput, reverse genetic analyses of gene function in fern gametophytes.

BMC Pl. Biol. 4:6. (online journal, see http://www.biomedcentral.com).

SALMI, M. L., T. J. BUSHART, S. C. STOUT and S. J. ROUX. 2005. Profile and analysis of gene expression

changes during early development in germinating spores of Ceratopteris richardii. Pl.

Physiol. 138:1734–1745.


STOUT, S. C., G. B. CLARK, S. ARCHER-EVANS and S. J. ROUX. 2003. Rapid and efficient suppression of

gene expression in a single-cell model system, Ceratopteris richardii. Pl. Physiol.

131:1165–1168.

SULLIVAN, K. F. 1988. Structure and utilization of tubulin isotypes. Annual Rev. Cell. Biol.

4:687–716.

VAUGHAN, K. C., L. G. HICKOK, T. R. WARNE and A. C. FARROW. 1990. Structural analysis and

inheritance of a clumped-chloroplast mutant in the fern Ceratopteris. J. Heredity 81:146–151.

VOGELMANN, T., A. BASSEL and J. MILLER. 1981. Effects of microtubule-inhibitors on nuclear migration

and rhizoid differentiation in germinating fern spores (Onoclea sensibilis). Protoplasma

109:295–316.

YOUNG, R. A. and R. W. DAVIS. 1991. Gene isolation with l gt11 system. Methods Enzymol.

194:230–238.

ZHANG, J. 2003. Evolution by gene duplication: an update. Trends Ecol. Evol. 18(6):292–298.


Documents

Characterization of Four Members of the Alpha-Tubulin Gene